info
Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Re: Comments in WebRender
22 views
Skip to first unread message
Kartikaya Gupta
Sep 28, 2018, 9:27:16 AM9/28/18
to nical...@gmail.com, Bobby Holley, Glenn Watson, Dzmitry Malyshau, Jeff Muizelaar, Alexis Beingessner, Nicolas Silva, Lee Salzman, sid...@mozilla.com, Kartikaya Gupta, Andrew Osmond, dev-te...@lists.mozilla.org
One meta-comment: the dev-tech-gfx mailing list is a good one for
these sorts of emails since most if not all people working on
gfx/webrender are on there (if they are not, they should be). I'm
adding the mailing list to the CC on this reply.
And one response to the IRC discussion I saw in the backscroll:
Dzmitry said that "It's just that in an ideal world, the semantics is
expressed *mostly* by proper types and naming, and comments helping on
top." - which I agree with, but with one caveat. The caveat is that
proper types and naming help much more if you already have a good
mental model of the high-level goals and architecture of the code;
without that the names are basically meaningless. Newcomers to the
code may not have this mental model, and so the comments help build
that and provide more meaning to the types/names being used. As a
hopefully concrete example, the name "ResourceCache" is obviously a
cache of resources, but somebody unfamiliar with the code would then
ask "what are resources?" and if that's not documented it's going to
be a painful process to figure it out. So +1 from me for adding
comments and getting to a deny(missing_docs) state, but obviously try
and make those comments as meaningful as possible.
On Fri, Sep 28, 2018 at 6:10 AM Nicolas Silva <nical...@gmail.com> wrote:
>
> Sounds good to me. My only ask is that if anything is enforced in the CI, it is also enforced in local developer builds.
>
>
> On Fri, Sep 28, 2018, at 1:28 AM, Bobby Holley wrote:
>
> I posted this on IRC, but jrmuizel and nical are offline, so moving to email. Please CC any other WebRender reviewers I'm missing.
>
> https://github.com/servo/webrender/wiki/Comments-in-WebRender
>
> That sound good to everyone? If so, I'd appreciate if everyone here enforce this during review.
>
> Thanks!
> bholley
>
>
these sorts of emails since most if not all people working on
gfx/webrender are on there (if they are not, they should be). I'm
adding the mailing list to the CC on this reply.
And one response to the IRC discussion I saw in the backscroll:
Dzmitry said that "It's just that in an ideal world, the semantics is
expressed *mostly* by proper types and naming, and comments helping on
top." - which I agree with, but with one caveat. The caveat is that
proper types and naming help much more if you already have a good
mental model of the high-level goals and architecture of the code;
without that the names are basically meaningless. Newcomers to the
code may not have this mental model, and so the comments help build
that and provide more meaning to the types/names being used. As a
hopefully concrete example, the name "ResourceCache" is obviously a
cache of resources, but somebody unfamiliar with the code would then
ask "what are resources?" and if that's not documented it's going to
be a painful process to figure it out. So +1 from me for adding
comments and getting to a deny(missing_docs) state, but obviously try
and make those comments as meaningful as possible.
On Fri, Sep 28, 2018 at 6:10 AM Nicolas Silva <nical...@gmail.com> wrote:
>
> Sounds good to me. My only ask is that if anything is enforced in the CI, it is also enforced in local developer builds.
>
>
> On Fri, Sep 28, 2018, at 1:28 AM, Bobby Holley wrote:
>
> I posted this on IRC, but jrmuizel and nical are offline, so moving to email. Please CC any other WebRender reviewers I'm missing.
>
> https://github.com/servo/webrender/wiki/Comments-in-WebRender
>
> That sound good to everyone? If so, I'd appreciate if everyone here enforce this during review.
>
> Thanks!
> bholley
>
>
0 new messages